Every time Sam Altman had to apologise publicly

Sam Altman, the man behind the leading artificial intelligence (AI) startup, OpenAI, is one of the most recognisable names in the AI industry.

Once president of the prestigious startup accelerator Y Combinator, Altman went on to co-found OpenAI in 2015 with Elon Musk and other leading technologists.

Under his leadership, OpenAI has developed ChatGPT, GPT-4o, and GPT-5, reshaping how people interact with technology and sparking both excitement and fear worldwide.

However, in several instances, Altman has publicly apologised, acknowledging mistakes, addressing missteps, and attempting to restore trust.

Here are five public apologies that shaped both Altman’s public image and OpenAI’s direction.

1.Altman admitted AI is replacing jobs, said sorry (July 2023)

In 2023, Altman sparked debate when he admitted in an interview with “The Atlantic” that AI was not just supplementing human labour, it was actively replacing jobs.

He criticised fellow tech leaders for painting an overly optimistic picture and bluntly said:

“Jobs are definitely going to go away, full stop.”

Altman later acknowledged the harsh reality of his words, apologising for the way he delivered the message.

2. Apology over harsh NDAs at OpenAI (May 2024)

In 2024, controversy erupted after reports revealed OpenAI employees were bound by strict non-disclosure agreements (NDAs) that threatened to revoke vested equity if they criticised the company.

The development came after the high-profile researchers, including Ilya Sutskever and Jan Leike, resigned.

Altman admitted he was unaware of the severity of the contracts and took to X (formerly Twitter) to say he was “genuinely embarrassed.”

He apologised, confirmed the clauses would be revised, and stressed that OpenAI had never actually clawed back anyone’s equity.

3. Scarlett Johansson and the “Sky” voice (2024)

One of Altman’s most high-profile apologies came after Hollywood actress Scarlett Johansson accused OpenAI of creating a chatbot voice that sounded “eerily similar” to hers.

Johansson revealed she had preiously declined an offer to voice ChatGPT, only to be stunned by how closely the AI resembled her tone.

Following the backlash, Altman issued a statement denying intentional imitation but apologised to Johansson for the misunderstanding and paused the “Sky” voice.

4. “I feel sorry for Elon Musk” (February 2025)

Altman’s feud with Musk took a dramatic turn in early 2025, when reports surfaced that Musk was leading a $97.4 billion bid to take over OpenAI’s nonprofit board.

Altman dismissed the attempt with sarcasm, even offering to buy Twitter (now X) in return. 

But during an interview at the Paris AI Action Summit, Altman’s remarks grew more personal.

He said Musk’s actions stemmed from “bottomless insecurity” and added, “I do feel for the guy.”

5. OpenAI’s open-weight model delay

In July, Altman announced a delay in the release of OpenAI’s open-weight model, citing safety concerns.

Originally scheduled for launch the following week, the model was postponed indefinitely for additional testing.

Altman apologised to the community, writing: “Sorry to be the bearer of bad news; we are working super hard!”

6. The GPT-5 “mega chart screwup”

Altman’s next embarrassing moment came during the launch of GPT-5 in August, when OpenAI presented misleading charts that exaggerated the model’s capabilities.

The visuals incorrectly suggested GPT-5 was far superior at avoiding deception than it actually was.

Altman quickly admitted the error, calling it a “mega chart screwup.” OpenAI issued corrections on its blog and apologised for what it described as an “unintentional chart crime.”